1 Introduction

The large alluvial plains (Indo-Gangetic plains) owe their evolution to the sediment generated and transported by the rivers originating from the Himalayan orogen. Aravalli–Delhi Ridge, which runs northeast to southwest, divides alluvial plains into two drainage basins (Singh 1996): the western alluvial plains (Punjab–Haryana) and the eastern alluvial plain (Ganga–Brahmaputra). A wide palaeochannel belt, the Ghaggar–Hakra channel is present in vast Indo-Gangetic Plains. This palaeochannel belt is deprived of the major drainage system, and it is mapped between the Ganga and the Indus encompassing the east and west sides of the Indo-Gangetic Plains, respectively (Yashpal et al. 1980; Valdiya et al. 2005, 2013, 2016; Singh and Sinha 2019; Chaudhari et al. 2021). Saraswati is the mythological river of the Himalayas and is also believed to have existed in this northwestern part of vast alluvial plains and was flowing through the parts of Haryana, Punjab, Rajasthan, Gujrat, and modern Pakistan. In the last couple of decades, many types of research have been carried out in the Indo-Gangetic alluvial plains, and their relation with the Himalayas and our understanding of various aspects has been enhanced significantly; however, the existence of the Saraswati River and its course is still unresolved, due to the paucity of the subsurface geological data of the river’s palaeocourse. Therefore, the workers are now taking more interest in this direction, and the work of Sinha et al. (2013), represented resistivity data along a buried palaeochannel of the Ghaggar River, wherein they inferred that there was a high energy wide braided river system which subsequently changed to a narrow channel seasonal river system. In a study, Singh et al. (1997) referred to two major tectonic phases, the older ~40 ka and the other at ~5–6 ka that caused the reorientation of several river systems. Recently, Singh et al. (2016) represented data from two drill cores from the buried palaeochannel of palaeo-Ghaggar. The study shows high-energy channel sand body deposit overlain by fine-grained, low-energy channel deposit, which is further overlain by reworked floodplain deposits. Based on the geochemical analysis, including the Rb–Sr isotopic composition, they established that sediments are derived largely from the higher Himalayas and once a well-established large river system existed in the region. The studied palaeochannel may be the palaeochannel of the Saraswati River (Chaudhari et al. 2021). In some relatively recent studies of buried palaeochannel of the Ghaggar River (Sinha et al. 2013; Singh et al. 2016), it has been inferred that the fluvial system setting changed from high energy wide network of braiding rivers to an incised river channel system and ultimately transformed into an ephemeral river system with small channels due to dwindling climatic conditions. The role of tectonics, however, coupled with climate may also have been invoked in some earlier studies (Singh and Sinha 2019).

The geochemical composition of siliciclastic sediments provides credible information on geological earth processes such as weathering and sediment transport, provenance, tectonic setting, etc. (Bhatia 1983; McLennan and Taylor 1983; Taylor and McLennan 1985; Wronkiewicz and Condie 1987; Cullers 1988; Feng and Kerrich 1990; Condie et al. 1992; McLennan et al. 1993; Garver and Scott 1995; Fedo et al. 1996; Nesbitt et al. 1996; McCann 1998; Singh and Rajamani 2001a, b; Tripathi et al. 2004; Singh 2009). The present study would provide insight into the provenance, weathering, and tectonic setting of the palaeochannel.

2 Study area

The study area lies ~13 km west of Kurukshetra (between Pehowa and Bhor–Saidan villages, 29.965313N; 76.699775E) (figures 1 and 2). It is bordered in the north by the Himalayan foothills, in the east by the Yamuna River, in the west by the huge Thar Desert, and in the south by degraded Aravalli Hills. Markanda and Ghaggar rivers flow towards the southwest (western alluvial plain), whereas Yamuna River flows towards the east (Indo-Gangetic alluvial plain) (figure 3). The climate is semi-arid with mean annual precipitation of around 450 mm most of which is received during the Indian Summer Monsoon (ISM). The study area experiences maximum and minimum temperatures of 45 degrees and 2 degrees Celsius, respectively. Vegetation is dominated by tropical dry deciduous in the northeastern region; the Siwaliks region is covered by tropical moist deciduous and the western region has tropical thorn forests.

Figure 1
figure 1

Map of India showing Haryana state, with further elaboration of the study area.

Full size image
Figure 2
figure 2

Google Earth map showing the location of the study area, yellow lines on the map are the state highways connecting the study area.

Full size image
Figure 3
figure 3

Simplified geological map of NW Himalaya along with major drainage and site location BS. The map is taken from Singh et al. (2016), based on the compilations taken from Webb et al. (2011), Yin (2006), Vannay et al. (2004), and Myrow et al. (2015).

Full size image

3 Methodology

The 6.5-m thick exposed sedimentary section of a palaeochannel was systematically sampled. A trench was made to collect samples, leaving the uppermost 0.6 m to minimise anthropogenic and surface contamination. A total of 62 samples were collected at 10 cm for sand-dominated horizons and 5 cm for clay-rich layers, out of which 36 representative samples were selected based on textural variations for geochemical analysis.

3.1 Textural analyses

Grain size analyses were carried out using a laser particle size analyser (Beckman-Coulter LS 13 320). For which, 10 g of sample was treated with different chemical reagents such as sodium acetate (1 N), hydrogen peroxide (30%), sodium citrate (0.3 M), and sodium dithionite combination to remove, carbonate, organic carbon, and iron-manganese coatings, respectively (Jackson 1956; Kunze and Dixon 1986). Finally, 1–2 ml of the well-homogenised samples were analysed in the laser particle size analyser.

3.2 Mineralogical and geochemical analyses

X-ray diffraction (XRD) analysis was done on well-crushed and homogenised powdered samples. The size and weight of the samples are <63 µm and 1–2 g, respectively. PANalytical Xpert3 instrument was used to find out the bulk mineralogical distribution. A homogenised powdered sample with a mesh size of around 200 microns was utilised for the analysis of the major elements. To make the palettes, a 6:4 ratio of sample to boric acid was compacted (2500 psi) using a Kameyo Powder-Press. Using a wavelength-dispersive X-ray fluorescence spectrometer, the pellets were examined (WD-XRF; PANalytical AXios mAX). In this technique, a programme for computers that uses the matrix calibration method transformed X-ray counts into concentrations (Franzini et al. 1972). For major oxides, the analytical precision was 5%, while the accuracy ranged from 2 to 10%.

ICP-MS (Agilent 7700) is used to examine trace elements, together with rare earth elements (REE). Milli-Q ultra-pure (18.2 MΩ-cm) water was used to make a solution. All the samples were digested by taking 30 mg (–200 mesh) sediment powder by using supra pure acids (HF, HCLO4, HNO3). Four solutions (10, 50, 100, 200, and 300 ppb for all elements) were prepared by 71A and 71B multi-element calibration standard solutions (Inorganic Ventures make) for the external calibration of the machine. Two USGS rock powder standards, namely the Green River Shale (SGR-1b) and the Cody Shale (SCo-1), were used for analyses. All the datasets were well below from 5% error with a good calibration curve. The geochemical analyses were carried out at Birbal Sahani Institute of Palaeosciences, Lucknow.

3.3 Optically stimulated luminescence (OSL) dating

The date of the sedimentary succession of the palaeochannel was established using an optically stimulated luminescence (OSL) dating technique since the sediments contain less organic material (Aitken 1998; Wallinnga 2002). OSL dating is a widely used method to determine the age of river deposits by estimating the time since the last time they were exposed to sunlight (Duller 1996; Jain et al. 2004; Cordier et al. 2010; Morthekai and Ali 2014). The OSL samples were collected from the scraped section along with the sediment samples. Samples were opened in the OSL lab under low-intensity red lighting. To determine the water content and dosing rate, the pipe's first 3 cm sediment sample was taken out and preserved. The sediment samples were processed with 1N HCl to remove carbonates and 30% H2O2 to remove organic debris to extract the quartz. The samples were then dried in an oven at 45°C after being cleaned with distilled water. The dried samples were sieved to get grains that were 90–150 µm in size. Heavy minerals and feldspar were removed from the quartz fraction using an isodynamic separator set at 0.5 A and 1.5 A (Porat 2006). To remove the outer alpha skin (~20 µm) and remaining feldspar, the cleaned quartz portion was scraped with 40% HF for 80 min. The samples were then cleaned of fluorides by being exposed to 12 N HCl for 30 minutes. For measurements, an automated Risø TL-OSL reader (TL/OSL-DA-20) (Bøtter-Jensen et al. 2010) is used. An on-plate 90Sr/90Y beta source with a dosage rate of 5.96 Gy/min was used to irradiate the samples. To calculate the annual dose, the conversion factor suggested by Adamiec and Aitken (1998) was used to assess the radioactive material concentrations. A highly pure germanium detector (HPGe) was used to measure the U, Th, and K concentrations. To achieve radioactive equilibrium, the samples are stored for 21 days after being enclosed in plastic boxes. Determination of the cosmic rays to the dosage rate is calculated using the average water content of 5–10% and the Prescott and Hutton (1994) technique. Four samples were collected to establish the chronology.

4 Result

4.1 Alluvial stratigraphy

The sediments are dominated by grey and micaceous sand and silt facies with current bedding, ripple marks, and planar laminations, and are marked as Markanda Formation (Younger Alluvium) in the lithostratigraphic column. Further, silty sand layers are intercalated with sand layers showing fluvial cycle influence in the palaeoriver channel (Thussu 1995). Based on the sediment texture and structures, the 6.5-m thick alluvial sequence is divided into nine units (figure 4). Stratigraphically, the base is not exposed. Unit 1 of the section is 1.6-m thick and is dominated by finer sand particles (96–98%) with only 2–3% of silt-size particles. Yellow-coloured oxidised sand layers were also observed in unit 1. Unit 2 is around 0.4 m in thickness and has about 18–25% of silt, 75–80% of sand, and 1–2% of clay. Unit 3 is again 0.4 m thick fine sand with 95–98% of sand fraction, while 2–3% of silt is present. Unit 4 is 0.3 m thick and has a much more fraction of silt than unit 3, which is below it. Silt is about 18–25% in this unit. Unit 5 is 0.6 m thick, is dominated by finer sand particles, and is again a sandy layer. Unit 6 is 0.8 m thick and is composed of 18–20% of silt, 4–5% of clay, and 70–80% of sand. Unit 7 is 0.2 m thick and is a silt-dominated unit with 10–15% clay, 70–80% silt, and 10–20% of sand-sized particles. Unit 8 is 0.2 m thick and again has more silt (60–70%) percentage than sand (15–25%); however, the clay percentage is lesser than in unit 7. Unit 9 is 1.65 m thick and is dominated by clayey silt, which has 10–15% of clay. Reddish brown silty clay of this unit is rich in calcareous white nodules. Above unit 9 lies ~0.35 m thick pottery-dominated horizon, mixed with black to light brown scree, which indicates human inhabitation and is considered as the cultural layer.

Figure 4
figure 4

Field photograph of the Bhor Saidan Palaeochannel section showing different terraces and the litholog with details of texture and sedimentary structure. Sample locations are represented as solid dots outside litholog (bottom to top).

Full size image

4.2 Texture analysis

The variation in the mean size value of the data reveals that samples consist of very fine sand to medium silt sediments having 10–16% clay mixed with medium silt samples (figure 5). Silt samples present in unit 9 of the sequence are bimodal and trimodal, poorly sorted, and mesokurtic to very leptokurtic in nature. The sediments from units 8 and 7 of the sequence are unimodal and finely skewed. Fine sand samples of units 1 and 3 of the sequence show well-sorted grains which are near-symmetrical and mesokurtic in nature. Sorting in this unit is attributed to the recycling and transportation of the sediments. The interrelationship of various parameters shows the unimodal to bimodal nature of sediments (table 1).

Figure 5
figure 5

Depth-wise distribution of the grain size data of the lithology present in the area. This figure shows the number of samples and percentage of the grain size distribution (sand, silt and clay) from bottom to top of the section. Silt and clay dominate the upper part of the section.

Full size image

4.3 Chronology

OSL dating has been done on four samples. The bottom-most sample from unit 1 has provided an age of 3.3 ± 0.2 ka. Second sample was taken from unit 6 of the section with silty sand lithology, which yielded the oldest date of the section, i.e., 11.2 ± 0.9 ka. The third sample was taken from the sand lithology of unit 7, which had given the age of 7.4 ± 2.4 ka. The top sample from unit 9 has given the age of 7.3 ± 0.7 ka (table 2).

4.4 Mineralogy and geochemistry

4.4.1 Mineralogy

X-ray diffraction (XRD) for 23 samples was used to scrutinise detailed mineralogy. These 23 samples represent the textural variations of the whole section. The bulk mineralogy of palaeochannel sediments mainly consists of quartz, K-feldspar, plagioclase, micas, and clay minerals (illite, chlorite, and montmorillonite) (figure 6). K-feldspar, with an abundance of quartz, is mostly present in the sand samples (figure 6a) and clay mineral illite is dominant in clayey-silt samples (figure 6b), showing the presence of water or humid climate after deposition (Keller 1962). The XRD pattern's variance in peak height may have been principally controlled by increasing quartz content (coarser fraction) in the samples; maybe it would have prohibited the other constituent minerals' peak heights from rising (figure 6a, b). Heavy minerals are significant provenance indicators and are also sensitive to weathering, transportation, deposition, and diagenesis (Morton 1985a, b). Heavy mineral assemblages of the palaeochannel mainly consist of tourmaline, zircon, garnet, staurolite, epidote, rutile, chlorite, and biotite. Garnet comprises the maximum percentage in the area followed by zircon, epidote, and tourmaline (figure 7). Zircon grains are of euhedral shape. The zircon-tourmaline-rutile index (ZTR index; Hubert 1962) was determined using the relationship ZTR Index = [(Z+T+R)/Non-opaque]. The ZTR index is 47%, indicating that the sediments were moderately weathered (figure 8).

Figure 6
figure 6

(a and b) XRD data for sand samples and the clay samples. Graphs clearly show the presence of quartz low and its peak is dominating another minerals peak.

Full size image
Figure 7
figure 7

Heavy minerals present in the paleochannel. Pictures clearly show sub-angular to sub-rounded crystals.

Full size image
Figure 8
figure 8

Diagram showing the percentage of heavy minerals in the paleochannel. The zircon–tourmaline–rutile index comes out to 47%.

Full size image

4.4.2 Geochemistry

Identification of provenance, weathering of the source area, and tectonics can be done through the chemical composition of the clastic sediments (Pettijohn 1972; Bhatia 1983, Bhatia and Crook 1986; McLennan and Taylor 1983; Wornkiewicz and Condie 1987; Fedo et al. 1996; Nesbitt et al. 1996; McCann 1998; Singh and Rajamani 2001a, b; Tripathi and Rajamani 2003). The major elemental geochemistry of the sediments is given in table 3. The concentration of SiO2 and Al2O3 in bulk sediments varies from ~50 to 85 wt% and ~10 to 18 wt%, respectively. Fe2O3 (~2–8 wt%) and MgO (~0.5–2.5 wt%) show good variance. MnO (~0.02–0.11 wt%), TiO2 (~0.2–0.7 wt%), P2O5 (~0.1–0.3 wt%), K2O (~1.8–3.5 wt%), and Na2O (~0.2–1 wt%) are in restricted ranges, CaO (~0.4–10.8 wt%) show large variation (table 3, figure 9). Cross plots (figure 10) of major oxides with Al2O3 show that there is a positive correlation with other major oxides such as TiO2 (r = 0.89), Fe2O3 (r = 0.97), K2O (r = 0.98), and MgO (r = 0.91), where the increasing trend of these elements probably points towards enhancement of clay minerals (Nagarajan et al. 2007a, b). A negative association with SiO2 (r = 0.93) and Na2O (r = 0.53) indicates that the sediments come from well-developed continental provenance. Enrichment of Al and Fe points towards the formation of clay minerals, mainly of 2:1 type (illite) (Keller 1970a, b). Depletion of Na more than K shows the alteration of plagioclase feldspar. Compared to Na, Ca is not depleted much; however, this may result from secondary carbonate formation demonstrated by the stepwise loss on ignition (LOI) values (table 3). The concentration of TiO2 is more in clayey silt (≤0.7) as compared to sands (≤0.03), suggesting less amount of phyllosilicate minerals in sands (Dabard 1990; Condie et al. 1992; Nagarajan et al. 2007a). Major REE and trace elements have a positive correlation with the grain size distribution. The amount of trace element values is given in table 4. Cr (r = 0.4), Co (r = –0.9), V (r = 0.9), and Cu (r = 0.8) have lesser values in the sediment samples compared to NASC and PAAS, but are associated with Al2O3 positively, indicating their association in clay-rich sediments (Wronkiewicz and Condie l987). Charts of the rare earth elements that have been chondrite-normalised (figure 11) show that the concentration of LREE (light rare earth elements) is much higher than HREE (heavy rare earth elements). The negative Eu anomalies are just like the values of UCC (McLennan et al. 2001). Values of HREE do not form a consistent or parallel pattern with the UCC, PAAS, or NASC (figure 11), which can result from the grain size effect (Taylor and McLennan 1985). Moreover, the concentration of HREE is less in sand-size samples, which may be credited to the quartz dilution effect (Götze and Lewis 1994).

Table 1 Textural analysis data of the sediments.
Full size table
Figure 9
figure 9

Grain size (sand, silt ± clay) distributions in the palaeochannel profile along with the litholog. Note that SiO2 and felsic components are more in the coarser mode (sand), whereas mafic components are enriched in the finer sediments.

Full size image
Figure 10
figure 10

Cross plots of major oxides with Al2O3 show that there is a positive correlation of major oxides like TiO2 (r = 0.89), Fe2O3 (r = 0.97), K2O (r = 0.98), MgO (r = 0.91), whereas SiO2 (r = –0.93) and Na2O (r = –0.53) show negative correlation.

Full size image
Table 2 OSL dating results of the sediment samples.
Full size table
Figure 11
figure 11

(a) Chondrite normalised samples are plotted to compare with UCC, PAAS and NASC. The pattern shows negative Eu anomaly. (b and c) PAAS and UCC normalised samples showing a similar pattern with positive Eu anomaly for one or two samples. Sources: chondrite (McDonough and Sun 1995), UCC (Rudnick and Gao 2003) and PAAS (Taylor and McLennan 1985).

Full size image

5 Discussion

5.1 Sediment maturity and source area weathering

Sediment suites with various chemical compositions are often produced by the hydrological differentiating processes that take place on river sediments in combination with source rock composition and chemical weathering (Nesbitt and Young 1989; Frallick and Kronberg 1997). As sediments become more mature, quartz replaces feldspar, mafic minerals, and lithic fragments in the sediments. As a result, the major elements Na, K, Ca, Al, Fe, Mg, and certain other trace elements are reduced, and SiO2 is increased in the bed load sediments. Some ratios, such as SiO2/Al2O3, Na2O/K2O, Fe2O3/K2O, and Fe2O3/SiO2, which serve as effective markers of sediment maturity, are impacted by this (Pettijohn et al. 1972; Herron 1988). SiO2/Al2O3 ratios rise together with sediment maturity, whereas Fe2O3/SiO2 and Al2O3/SiO2 ratios fall. The ratios Na2O/K2O and Fe2O3/K2O help in predicting the stability of feldspar and minerals bearing Fe and K. The sediments of the study area are plotted between log(Na2O/K2O) vs. log(SiO2/Al2O3) (Pettijohn et al. 1972), which show that several of the sediments are arkosic (figure 12). Units 9 and 7 of the section is clay enriched and falls in the arkosic field, and the extremely low Na2O/K2O ratio due to in situ chemical weathering of clay enriched units 9 and 7 may be the reason for shifting the samples towards the arkosic field. Sand sediments fall in the litharenite segment inferring more physical weathering or less alteration of feldspar (Singh 2010). Graph Fe2O3/K2O (Heron 1988) shows that arkosic arenites are grading into shale with wackes in transition (figure 12) units of the section with high silt and clay percentage are having high Fe2O3/K2O ratio and the negative relation of SiO2 with Fe, Mn, and Ti point towards the weathering of the sediments. This can be attributed to the mature and recycled nature of the sand grains (Bhatia 1983) as well as the formation of illite clay minerals.

Figure 12
figure 12

Plot of sediments on the geochemical classification diagrams after Herron (1988) (a) log(Fe2O3/K2O) vs. log(SiO2/Al2O3) and (b) log(Na2O/K2O) vs. log(SiO2/Al2O3) (Pettijohn et al. 1972). Diagrams are clearly showing that the sediments which are dominated by sand lies in arkose field, silt dominated part lies in wacke field and clay dominated part lies in shale field.

Full size image

The value of Pb corresponds to the grain size effect (Vital et al. 1999a, b) and, therefore, is found higher in the clay-rich sediments than the value of PAAS and the value is less in the sand-rich sediments. A similar pattern is observed with V, Cr, U, Sc, Y, and Th, clearly showing the quartz dilution effect (Götze and Lewis 1994) in the sand-rich samples, affecting the absolute values of all other elements. LREE are positively correlated to the Al2O3, specifying that these are dominantly related to the clay minerals (Taylor and McLennan 1985). UCC and PAAS normalised plots of sediments show variation in the Eu anomaly values. Coarser grain size sediments have positive Eu anomalies, while clay size sediments show negative Eu anomalies; it might happen if the source location experiences less chemical weathering (Middelburg et al. 1988) as well as retention of feldspar in the sediments responsible for positive Eu anomaly. Moreover, the sand samples have not undergone much mineralogical sorting could also be the reason for the positive Eu anomaly (Bhatia 1985; Singh and Rajamani 2001a, b).

During weathering, smaller cations like Sr, Ca, and Na, deplete in the initial stage; however, the larger cations (Al and Rb) remain immobile, rather enriched in due course of time (Nesbitt and Young 1982, 1984; McLennan et al. 1993; Fedo et al. 1995). To determine the degree of weathering, CIA is commonly used (Nesbitt and Young 1982), where CIA = {Al2O3/(Al2O3 + CaO* + Na2O + K2O)} × 100 (*CaO in silicate fraction only). In the current study, CIA values range from 63 to 77. Al2O3, Na2O, and CaO* are displayed in the A–CN–K ternary diagram to assess the migration of the elements throughout the development of chemical weathering. It has been observed that the sediment samples plot over the plagioclase–K-feldspar line. The A–CN–K plot shows that most of the fall at the A–K edge as they approach the illite composition denotes moderate weathering (figure 13). When analysing the constitution of the source rock, the A–CN–K ternary plot excels because it may be projected backward and parallel to the A–CN line of weathered samples up to a position on the feldspar connect (Fedo et al. 1995; Tang et al. 2012). Siliciclastic sediments of the current research are found to be derived from granite as a probable source and to extend up to the illite stability zone beside a trend line parallel to the A–CN axis. Thus, the diagram implies that samples originated from granitic rocks as the source is influenced by weak chemical weathering (Madhavaraju et al. 2016). Moreover, figure 13 also shows that the major Himalayan rivers' sediments plot in the zone of weak weathering (CIA = 55–65), whereas the Ghaggar sediments show an intermediate degree of weathering (CIA = 69). Overall, the intermediate CIA values (63–77) for the palaeochannel sediments indicate semi-arid, water-starved granitic sources. Further, the textural attributes support physical weathering with a tectonic control.

Figure 13
figure 13

A–CN–K ternary plot (Nesbitt and Young 1982) showing chemical maturity of paleochannel sediments. Also shown for comparison are modern river sediments from Himalaya, average values for world river sediments (Li and Yang 2010), UCC, and PAAS.

Full size image

5.2 Provenance of sediments

In the current study, the LREE values are much higher than HREE and do not form a consistent or parallel pattern with the PAAS or NASC (Taylor and McLennan 1981), which may result from the grain size effect (Taylor and McLennan 1985). In our samples, LREE are more positively correlated to Al2O3, indicating that these are dominantly connected to the clay minerals. It is intriguing to observe that coarser sediments normalised with PAAS and NASC show positive Eu anomaly with lower HREE abundance, probably indicating the dominance of physical weathering over chemical weathering in the source region. As opposed to that, the samples of the sandy silt and clayey silt texture show negative Eu anomaly, which could result from plagioclase weathering at the deposition site. There is a positive association between LREE and Th, indicating felsic lithologies may have contributed (McLennan et al. 1980). The La/Th value ranges between 2.30 and 4.01, usually similar to UCC and PAAS, respectively, clearly indicating the grain size bias where coarser sediments are associated with granitoid and relatively finer sediments are of intermediate composition. A positive correlation is shown by REEs (table 5) to some extent with the other major oxides like MgO2, TiO2, and Fe2O3, however, they show a negative correlation with trace elements like Cu, Co, Cr, and U, showing the contribution of REEs from mica biotite and chlorite minerals (Taylor and McLennan 1985; Bauluz et al. 2000). La–Th–Sc ternary plot (Singh 2010) has been drawn to infer the provenance (figure 14). All the sediments fall in between the composition of granite and granodiorite.

Table 3 Major oxide data for the sediment samples of the palaeochannel, along with the grain size data. Variation in LOI and CIA is also presented.
Full size table
Figure 14
figure 14

La–Th–Sc ternary plot after Singh (2010) showing the composition of the source rock lies in between the granite and granodiorite.

Full size image

We have also applied the discriminant function scheme (figure 15) of provenance estimation proposed by Roscher and Korsch (1988). In the current study, sediments fall in the P4 part with a little shift towards the P1 part; this could be the result of the grain size effect (Whitmore et al. 2004). The shifting of results seen in this study is relatable to the study carried out by Roscher and Korch (1988) on sedimentary rocks of the Ordovician and the Silurian greywacke having SiO2 > 70%, where coarser sediments fall in P4 and finer sediments fall in P1 and P2 fields. The greywackes used in Roscher and Korch's (1988) study may be produced by the successive recycling of older sediments (Wyborn and Chappel 1983) and the composition of quartzose sedimentary sources can most probably be derived from varied sedimentary litho units and detrital components of different crystalline rocks. Similar results have been obtained in our study, as the sediments we have taken are also reworked having SiO2 > 70% and probably produced by successive recycling of older sediments.

Figure 15
figure 15

Discriminant function 1 against discriminant function 2 variation diagram. Fields after Roser and Korsch (1988), wherein F1 = −1.733TiO2 + 0.607Al2O3 + 0.76Fe2O3 − 1.5MgO + 0.616CaO + 0.509Na2O − 1.224K2O − 9.09, and F2 = 0.445TiO2 + 0.07Al2O3 − 0.25Fe2O. Provenance fields: (P1) mafic igneous provenance, (P2) intermediate igneous provenance, (P3) felsic igneous provenance, and (P4) quartzose sedimentary provenance.

Full size image
Table 4 Trace and rare earth element (REE) data (ppm) for the palaeochannel sediments.
Full size table

Heavy-mineral assemblages of terrigenous sediments also help in establishing the source of sediments (Carroll 1953; Crook 1968; Cleary and Conolly 1972; Colin et al. 1993; Oliva et al. 1999; Thomas et al. 1999; Horbe et al. 2004; Van Loon and Mange 2007). In the present study, the ZTR index is 46%, indicating that the sediments are moderately weathered and additionally corroborated by the CIA readings. The presence of garnet and epidote also shows the unweathered nature of the sediments (Hester 1974). Zircon grains present are of a euhedral shape and are derived from acidic igneous rocks (Basu 1985). All the grains are mostly sub-rounded to rounded showing their involvement in cyclic transportation (Pettijohn et al. 1973). The mixing of results led us to plot sediment bivariant plots (figure 16a, b, c, and d) (Singh 2010). Values of the Tethyan Sedimentary Series (TSS), High Himalayan Series (HHS), and Lesser Himalayan Series have been taken from Richards et al. (2005) and Siwaliks have been taken from Sinha et al. (2007). The above plots indicate the sediments mostly deviated towards Siwaliks. Siwaliks are formed from LHS and HHS (Sinha et al. 2007); therefore in the current study, the samples are also showing mixed geochemistry of the Himalayan ranges.

Figure 16
figure 16

Sediment bivarent plots after Singh (2010). Values of Tethyan Sedimentary Series (TSS), High Himalayan Series (HHS) and Lesser Himlayan Series have been taken from Richards et al. (2005) and Siwaliks has been taken from Sinha et al. (2007). The above plots indicate the sediments are mostly deviated towards Siwaliks.

Full size image
Table 5 Correlation coefficients (r) from correlation matrix obtained with mineralogical and geochemical data of palaeochannel sediments.
Full size table

5.3 Tectonic setting

Any sedimentary basin's tectonic context can be distinguished based on geochemistry (Bhatia 1983, 1984, 1985a; Roser and Krosch 1986; McLennan and Taylor 1991; Graver and Scott 1995). To ascertain the tectonic setting, discrimination diagrams (figure 17) suggested by Roser and Korsch (1986) and K2O/Na2O–SiO2/Al2O3 relationship diagrams (after Maynard et al. 1982) were applied. These illustrations imply that the samples were laid in a passive margin environment typified by mineralogically mature (quartz-rich) sediments probably brought by rivers to the deposition site. The northern Indian passive margin consists of the Higher and Lower Himalayas and a small part of the Indian craton and their sedimentary cover (Baud et al. 1996). The examined samples are located in the P4 field in the diagram showing a reworked orogenic landscape (quartzose sedimentary or granite-gneisses origin source region like PM) that was settled under the passive margin regime. The heavy mineral assemblages (discussed earlier also) clearly show that the sediments come from all parts of the Himalayas, including the Siwaliks, the lower Himalayas, and the upper Himalayas.

Figure 17
figure 17

Tectonic discrimination diagram (a) Roser and Korsch (1986) and (b) Maynard et al. (1982) samples plot in passive margin field (PM). ACM = active continental margin field, ARC = continental arc sediments plot in (PM) field. A1 = arc basaltic and andesitic detritus; A2 = evolved arc setting felsitic plutonic detritus.

Full size image

5.4 Depositional environment and climatic changes

Depositional environment of the sediments is well correlated with the climate change of the area. Unit 1 has current beddings and ripple marks, indicating a palaeochannel of a flowing river. The lower part of the section, i.e., units 1, 2, 3, and 5 containing sand-size grain particles are arkosic and are subjected to less chemical weathering as compared to the upper part of the section, i.e., units 6, 7, 8, and 9, comprising silt and clay, showing a much more matured part of the section. Units 7 and 8 have poorly sorted nature of the sediments with a high kurtosis value denoting that the depositional environment has less energy. Poorly sorted nature of sediments can be attributed to the in-situ chemical weathering condition of the depositional area. Overall textural characteristics show fining upward sequence and also point towards decreasing depositional energy conditions. Negative Eu anomaly similar to PAAS and UCC of the average sediments indicates some cratonic source of sediments. The positive correlation of REEs and major oxides like Al2O3, TiO2, K2O, MgO, and Fe2O3 shows that REEs are indicating towards the granitic source rock. The value of CIA is intermediate (63–77), inferring intermediate weathering of the sediments at the source area because of a higher rate of physical weathering over chemical weathering. The above data and OSL results reveal that the lower part comprising unit 1 is the youngest dating back to 3.3 ± 0.2 ka, which can result from an incision by some sudden flow of water resulting from some instant weather change. The upper part comprising unit 6 is the oldest (11.2 ± 0.9 ka) and units 7 (7.4 ± 2.4 ka) and 9 (7.3 ± 0.2 ka) are relatively mature in nature.

The texture, mineralogy, and geochemistry of sediments invoke that they are mature (mineralogical constraints), recycled (sub-rounded to rounded grains), and physically weathered (arkosic). The sediment terrain is felsic in nature, having high-grade metamorphic rocks (heavy mineral assemblages), most probably granite or granitoid. Besides, the sediments are derived from a high altitude and arid climatic area. The climate change can be attributed to the chemically matured unit 6 dating back to 11.2 ± 0.9 ka with CIA values ranging from ~69 to 70.5, with less percentage of Al2O3 and more of SiO2 percentage, which can be due to low depositional energy with dry climatic conditions, whereas units 7 (7.4 ± 2.4 ka) and 9 (7.3 ± 0.2 ka) have CIA values of ~75–77, more percentage of Al2O3 and less of SiO2 percentage showing warm and humid climate. These phases correspond well with the records of lake proxies of the Ganga plain, where 14–12.5 ka and 11.5–10.5 ka are marked by cool and dry climate (Sharma et al. 2004; Chauhan et al. 2015) and period of 8.5–6.4 ka (Trivedi et al. 2013; Saxena et al. 2015) has a warm and moderately humid climate. The incision at 3.3 ± 0.2 ka in the present study is matched with the palaeoflood events in the Ghaggar–Hakra plain of Punjab around 3.9 ka (Singh et al. 2021). Our geochemical data of the exposed section is very much similar to the one of the drill core data from Singh et al. (2016) and confirms the existence of a large river.

6 Conclusions

This palaeochannel implies that the river which brought the sediments is reworked and mostly from Siwaliks. It can be correlated as the sediments are reworked and matured, derived from the recycled orogenic terrain like granite-gneisses or quartz sedimentary provenance. Hence, the sediments that this river has brought may have faced cycles of transportation. The alternate occurrence of silty clay beds with sand beds indicates a change in the climatic condition. Tectonic setting, provenance, and source area weathering of these sediments clearly show that sediments are from Higher Himalayas and Siwaliks, and are influenced by the change in the climatic conditions.